DE3030255A1 - METHOD FOR TRANSMITTING WORDS AND MESSAGE TRANSMISSION SYSTEM FOR ITS IMPLEMENTATION - Google Patents

Description

LiPL.-ING. J. RICHTER.-ING. F. WERDERMANN Q.

• .f? a. .TE 'NT Λ N WA LT E

REINS. REPRESENTATIVE AT THE EPO -PROFESSIONAL REPRESENTATIVES BEFORE EPO · MANDATAIRES AGREES PRES L ¹ OEB D-2OOO HAMBURG 36 NEW WALL IO

ST (O 4 O) 34OO4S / 34OOSe TELESRAMS: INVENTIOS MAMBURS

Wdm / Wa

YOUR CHARACTER / YOUR FILE OUR CHARACTERISTICS / OUR FILE JJ. 80 3 ^{OATUM / DATC} β. 8. 8y

PATENT APPLICATION

PRIORITY: according to U.S. registration Serial No. 66.221 of August 13, 1979

TITLE: Method for the transmission of words and message transmission system for its implementation,

APPLICANT: NATIONAL SEMICONDUCTOR CORPORATION, 2900 Semiconductor Drive, Santa Clara, Caliph. 95 051 /V.St.A.

INVENTOR: Philip Marshall Roybai,

1111 Pippin Greek Court, San Jose, Calif. 95 120 /V.St.A.

130015/0722

The invention relates to a method of transmission of words, each of which is composed of characters encoded in a particular binary digit code are. The invention also relates to a message transmission system for carrying out this method.

A compression of the data to be transmitted is very desirable because it reduces the bandwidth or the transmission or transmission time or save on both. There has therefore been no lack of efforts to reduce the transmission time and compression to influence in different ways.

U.S. Patent 3,334,335 mentions a system in which a word is received in a language and is generated as an output value as a function of the word input of an address. The patent itself deals with a system / in which a second or foreign language word is output. It describes also systems in which a memory stores each word, an input word with the stored word compared and a signal generated if a match is found. This signal is subsequently used to determine the address of the correct stored word. The patent deals with the display of the Matching or mismatching the words.

In U.S. Patent 3,717,851 the processing is more densified Data described. In doing so, compressed data in the form of codes of various lengths are used. These codes

130015/0722

have attachments representing the length, which themselves are of different lengths and are coded. The resolutions with different Lengths are decoded by small, fast, associative memories intended only for searching, which have a matching provide the indicating signal as an address for a second memory, which has memory elements of the usual type. The output from the last main memory can be a base address for yet another memory and a display of the contained in the code word of different lengths · remaining bits.

Furthermore, a computer system is described in U.S. Patent 3,400,380, in which digits or 4-bit data representations are converted into a character or 8-bit format.

Further examples of conversion systems are found in U.S. patents 3,229,047, 2,662,347, 3,778,617 and 4,032,913.

A substantial part of the vocabulary to be used in messages is stored in a main read-only memory (ROM) Addresses stored that are stored by a so-called hashing algorithm. If the to be transferred Words of a message are received, each word is transformed into an address by the algorithm. The word can be stored at the address or the address can be attached to the end address of the word via a temporary read-only memory (ROM) in the main ROM. The word stored at the end address on the main ROM is combined with the word of the

130016/0722

righs compared. If identity is established, the Address transmitted. At the receiving station the address is retrieved and applied to a ROM which is the same Contains addresses and the same words stored at such addresses as the main ROM at the sending station. the Message is therefore put back together word for word.

If identity not established during transmission or a transmission of a special kind is to take place, a special character is formed and transmitted, which indicates that the subsequent broadcast or transmission is a character-by-character transmission and not an address. Provisions can be made for the transmission of special endings and special characters.

The communication system and method includes the use of read-only memory storing a vocabulary of words to be transmitted is stored at different addresses. The addresses are separated from the words by a Mixing algorithm derived. When a word is to be transferred, the mixing algorithm supplies an address. It can an intermediate address can be used in a buffer to save time and increase density of storage to obtain. The words stored in the memory are captured by accessing the final address, output or read out and compared with the word to be transmitted. If Identity is shown, the address will be

130015/0722

ORIGINAL INSPECTED

transfer. When there is no identity, the word becomes. Transfer character by character. By using a memory which contains a large majority of the words eligible for transmission can be a considerable Achieve data compression during transmission.

At the receiving station the address is received and a Memory supplied in which the same words and the same addresses are stored as in the memory at the Transmitting station. The stored word is read or output and the message is consequently reassembled. Words that have been transmitted character by character are also put together in this way.

The following are the invention, its advantages and possibilities explained in more detail, for example, for their further configuration with reference to the drawings. Show it

1 shows a basic block diagram of a transmitting station according to the invention;

2 shows a block diagram of a receiving station according to the invention;

3 shows a schematic illustration of part of a memory for explaining the understanding the invention;

4 shows a group of tables for explaining the application code;

5 with FIGS. 5A, 5B and 5C to explain the mode of operation of the device from FIG. 1 appropriate flow scheme;

130015/0722

6 shows a schematic representation of the storage two words in memory to explain a problem with addressing Words that have similar or nearly identical addresses;

Fig. 7 is a schematic diagram to assist in explaining how indirect addressing can increase the effectiveness of the working method and

8 and 9 examples of the advantages or savings through the method and the device according to the invention compared to ordinary ASCII (American Standards Code for Information Interchange) are achievable.

Referring to Fig. 1, there is an input device 10 which may be of any conventional type, for example one which provides an output signal according to ASCII (American Standards Code for Information Interchange), connected to a register 12, which this output signal records. The register 12 can be a register inside a digital computer 14. The ASCII signal code contains two bytes of four binary digits or bits each. Therefore, the connection between the input device 10 and the register 12 be an eight-wire cable. In this embodiment, a detection circuit 16 is inside the computer provided; but it could also be arranged outside of it. The detection circuit receives from the Register each ASCII character again when entered over an 8-bit cable. The detection circuit 16 may be a of the known decoding circuits, e.g. one

130015/0722

which contains one or more gate trees, or it can also be arranged by programming. If there is a gap "or a gap is detected by the circuit 16, activates a line P, which is used as a programming device for the address generation, the program for the address generation leads accordingly in the mixing algorithm as described in more detail below. The program can be stored in the internal memory of the computer 14 get saved. As a result of the execution of the program, each of the ASCII characters is in turn from register 12 read out (without being deleted) and, as indicated symbolically by the connecting line between register 12 and the Programming device 18 is illustrated applied to the latter; in this way the whole word is made up of different ASCII characters can be converted into a 16-bit start address. The start address is determined by the Programming device 18 a programmable auxiliary read-only memory (PROM) 19 and the final address stored at the start address in the auxiliary PROM is read out and fed to a comparator 22 and an AND circuit 24.

In the meantime, the comparator also receives from the register 12 each character when it is the programming device 18 for the address generation is supplied. The word delivered from the memory is then compared by the comparator 22 with the word received from the register 12 compared. if

130015/0722

the comparator indicates identity, receives the AND circuit 24 a signal from the comparator with a corresponding display via a connection to one of its outputs and enables the 16-bit address that is now may have been stored in a suitable register 30 is fed to a transmitter 28 for transmission. The registry 12 is then cleared for the next word. The transmitter may use a suitable temporary storage device such as E.g. a disk or the like included in order to transfer the characters, either bit for bit or one byte at a time, to accumulate.

If there is no identity, the address can be stored in memory either incrementally or up to the next EOR symbol (end of recording symbol) are advanced, then the next word can be read out and compared for identity and so on. As below explained in more detail, the successive words are often formed from the same root. The successive readings or outputs continue until identity is indicated or a certain number of samples are performed.

If the comparator 22 is still after the maximum number of samples indicating non-identity becomes a programmer 26 for the transmission character by character, which are also stored in the internal memory of the computer 14 can be activated to display the characters from register 12

13001S / 0722

subtract, either the single character that is in the register or consecutive characters, until the word stored in the register is extracted and fed to the transmitter by the programming device. In this case, the programming device can generate a special symbol which indicates that the next program is being broadcast for a reading character by character at the receiving point is intended, which means that at the receiving point another character is generated at the completion of the word to indicate that the character is occurring character by character The broadcast has ended. When all of the characters from register 12 have been output for transmission, the register is again released.

When the recognition circuit 16 receives certain characters recognizes as true, it can immediately activate the programming device 26 for a transmission character by character, but only for the transmission of the individual character in the register 12, and without this for the generation of the to generate the characters indicating character for character transmission, but with the addition of certain digits to a Complete an arbitrary 16-bit character that can be considered a pseudo-address. Such a pseudo-address can be recognized at the receiving station. At the same time the detection circuit 16 does the detection of a certain character as true, the display prevents the generation of the general address program by the programming device 18.

130016/0722

The transmitter 28 can like the 16-bit words it receives previously described, save for one or more transmissions and then either bit for bit, or in the form of bytes, or transmitted in the form of words, as appropriate in view of the transmission or transmission system used is to save bandwidth and reduce transmission time to compress.

FIG. 2 shows a receiver 40 for the transmission coming from the transmitter 28 of FIG. The 16-bit characters are is entered in a register 42, character by character, as soon as they are received. The 16-bit characters can be taken from the register 42 an auxiliary memory 46 and from there a programmable one Read-only memory 48 are fed. The memory 48 has the same words stored in the same addresses like the main read-only memory 20 of FIG. 1. The words from the read-only memory 48 of FIG. 2 can be read out one ASCII character for itself / i.e. one character with 8 bits or one byte. The readout time may be slower than the time for the address characters to be received by receiver 40. Therefore, the Auxiliary memory 46 can be used to provide temporary storage, each address from auxiliary memory is supplied after the word addressed by the next preceding address has been read from the PROM 48.

The output signal from the programmable read-only memory 48 can be fed to an output 52 via an OR circuit 50

13001S / 0722 COPY

such as a printer. The OR circuit 50 also receives 8-bit characters from a programming device 54 for receiving character by character, activated by an output signal from a detection circuit 56 which is switched in such a way that it reads the addresses or pseudo addresses coded with 16 bits from the register 42 receives. The recognition circuit 56 can thus ensure the recognition of certain specially coded characters, and these are translated by the programming device 54 and the OR circuit 50 as single characters in an ASCII code forwarded.

The working method should be clear from the above, but it should be summarized again as follows: A message to be transmitted is entered character by character at input 10. Every word is in the register 12 saved. If the characters are not recognized, either the special characters or a character gap, the register receives the word until a gap occurs and is recognized. When the gap is detected, the mixer programmer will 18 is activated for address generation and a 16-bit address is generated. The address becomes the auxiliary memory 19 and the programmable read-only memory 20 supplied. A word in ASCII code is read from memory 20 and entered into the comparator 22. This word is compared with the word read out from register 12. If identity is indicated, the address applied to memory 20 will be that stored in register 30 was, passed via the AND circuit 24 to the transmitter 2 8 and the

130015/0722

Ve;. ··. ■

Address is being sent. If not, the next word is read. When identity is achieved, that becomes the one in question Address sent. If not, the process will be indefinite Repeated maximum time of trial operations.

If after these samples have been carried out still no identity is determined by means of the comparator 22, the Programming device activated for broadcast character by character. The 8-bit characters are sequentially extracted from the register 12 read, transformed into 16-bit characters, of which 8 bits each correspond to the usual ASCII code, and through the transmitter 2 8 sent until the gap is read out by the programming device of the character for character transmission and the register 12 is empty or only contains gaps or zeros. At this point the programmer ensures 26, that all registers and circuits have been reset and the process can begin again.

When the recognition circuit 16 recognizes a special character, the programming device for address generation can be activated in order to store a special address in the memory 20 address in which the special character appears in the ASCII code. The ASCII code is then made with the character compared to register 12, and if identity is displayed, register 30 routes the 16-bit address via the AND circuit 24 into the transmitter 28 for transmission.

For the description of the storage process, it is to be assumed that experience has shown that the language vocabulary of a person

130015/0 7 22

contains an average of about 10,000 words and that your reading vocabulary is between 25,000 and 50,000 words. Therefore, when using 16 data bits (i.e. a 2-byte word or character) a system can also indicate the address of each word that is likely to be transmitted. If a special If vocabulary is desired, space for additional words is available or the vocabulary can be modified. Each word is stored in a large read-only memory (ROM). It can be beneficial to provide a programmable read only memory (PROM) such as memory 20, so that any desired changes to the Vocabulary or addresses are easy to implement. Therefore, each of PROMs 20 and 4 becomes 8 of FIGS. 1 and 2, respectively programmed to store the desired words. A typical English word usually requires five Up to ten characters, each character in the size of one byte, for representation when using ASCII. Therefore the Using the ROM as described, achieve a data compression factor of 2 to 4 when sending.

If necessary, numeric data can be transmitted as ASCII, or it can be a special one-off Character preceded, which indicates that what follows is a character for character transmission, and it can then the character for character transmission are terminated by the same or a similar special unique character.

The tables of FIG. 3 show various words stored in the PROM (or 48). The addresses can be saved as

13001 S / 0722

hexadecimal characters appear. Each hexadecimal character can have one of the values 0-15. Place the letters A-F corresponding digits for the value 10-15. In other words: Each hexadecimal character corresponds to a 4-bit digit, and the 4 characters of the address shown in FIG. 3 correspond to a 16-bit or 2-byte address. In Fig. 3 is The memory is shown in a completely schematic manner, usually the addresses occur in blocks each 'a multiple of 2, whereas here a memory length 23 is shown. The symbol or sign EOR (end of record) is used in a well-known manner and denotes the end of the Word. By using the symbol EOR, the words can be stored in a "condensed" manner, whereby the Storage density is increased beyond the use of a line for each word. The addresses are in hexadecimal Characters specified, namely four such characters or 2 bytes for an address. It should be noted that each address is refers to a place where there are at least 7 binary digits to provide the ASCII code for the character at that address.

The mixing algorithm is used to convert a word into an address which, in the case of this exemplary embodiment, corresponds to 4 hexadecimal characters. Any hexadecimal character can be converted to or viewed as a 4-bit number.

Before trying to put the presented word in 'the address in a hexadecimal code by the mixing algorithm

130018/0722

- # - '■' ·· "■■■■ -

walk, certain preparatory steps are taken. If the characters are sent to register 12 in ASCII code, recognizes a suitable detection circuit, either sub-circuit 16 or a separate circuit may, one of the punctuation marks shown in column 2 of Fig. 4. If such a character is recognized, it leads the computer a special memory, which is not shown, or something equivalent, namely a part of the PROM 20 at the appropriate address as indicated in column 1 of FIG. For example, the delimiter (/) is as the word DDD8 consisting of four hexadecimal characters is given. This hexadecimal word is then the 16-bit character which addresses the memory 20 * The word entries are then pulled out and introduced into the comparator 22. If Identity is shown, as it should be, it will Word DDD8 passed through the AND circuit 24 to the transmitter. In the same way can if any of the numbers 0-15 inclusive or -1 to -16 inclusive is recognized, the corresponding in column 3 for the numbers listed in column 4 specified hexadecimal word or the word specified in column 5 for column 6 numbers as the address be used. When a binary number is recognized or when a character is recognized that is not among those in the columns 4 and 6 falls, the calculator generates a number such as EE20 to indicate that the next byte • is a binary number, or like EE3O, to indicate that the next two bytes are binary numbers, etc., up to EE70, indicating that the next 6 bytes are binary numbers as indicated in columns 7 and 8. Possibly

130015/0722

these numbers can be sent one after the other in ASCII form, as described below, where EE8O or EE9O precedes or follows the broadcast. If the bytes to be followed are ASCII codes, this will be hexadecimal Word EE8O sent. In the same way, the hexadecimal word EE9O indicates that the following bytes have addresses accordingly are the mixed code according to the invention.

As for the shuffling algorithm itself, assume that a word is entered into register 12. The word or the combination of characters can be of the type indicated in column 10, and is then recognized and inserted into the appropriate The word given in column 9 is encoded with four hexadecimal characters. There can be additional words in the out four-character hexadecimal words are coded as indicated in column 9 below, starting with EEO2 through EEFF are. In other words, there will be as many words in the four hexadecimal character words EEO1 to EEFF coded as appropriate for the respective types of a shipment to be processed.

If none of these words are recognized, the machine will begin scanning from the character on the left, if it is written in English, looking for a consonant. When there are no consonants the code is FFFE, as with the last hexadecimal Characters indicated in column 11. The same encoding occurs when a consonant is found that is not is contained in column 12, such as the letter J or χ

130015/0722

On the other hand, if one of the consonants listed in column 12 is encountered, it becomes the one listed in column 11 Corresponding hexadecimal characters are used. The scan is then with respect to another consonant continued. If none is found, an F is entered. The scan is then for a third consonant continued in column 12. If it is not found, an F is entered again. If three consonants are found, only the last four letters of the word are taken into account. The last four letters are compared with the combinations listed in the table in column 14. If a match is found, so the hexadecimal character in column 13 that corresponds to the character in column 14 becomes the fourth character the address comprising four hexadecimal characters is used. The task can be broken down into a sequence of tasks in which the last four letters are adapted to the last four sequences of column 14, and if none If a four-letter match is found, the machine searches for three matching letters, and so on. If none Adaptation of the suffix as listed in column 14, is found, the corresponding hexadecimal character is an F as indicated in column 13.

In the operational or flow diagram of Fig. 5 are the manner in which each word is processed as well as the merging algorithm illustrated. The program begins as shown in Figure 5A above when the calculator writes a word (including

130015/0722

a punctuation mark) / for the register 12 is supplied. A comparison is then made to determine, whether the word is contained in column 10 of FIG. If so, the word is used as the corresponding four hexadecimal Word from column 9, Fig. 4, comprising characters. It then supplies the address for the memory. This code is used for the address, stored for transmission or sent, and the next word can be entered afterwards as indicated in the working diagram. If the word is not found in column 10 of Figure 4, a trial operation is initiated to determine if the word is a punctuation mark from column 2 of FIG. If the answer is yes the "word" from the corresponding entry in column 1 of Fig. 4 is coded and this code is used for the address, before considering the next word. If the answer is no, the word must be tested to determine whether it is a punctuation mark not found in column 2. If the answer is yes, go ahead the program advances to a point labeled W in the flow chart of Figure 5A. If the answer is no so the program asks back whether the word is a number between -16 and +15. If this is the case, it will be used of column 3 or 5 corresponding to the number contained in column 4 or 6 of FIG. The resulting Code is used for the address. If the answer is no, the question is whether the word is an out of range number Is -16 to +15. If this is the case, it will be listed as EE2O to EE7O and then a number in binary form, as in the entry

130015/0722

coded in column 7 according to the information in column 8. The resulting encoding is either sent or entered for temporary storage for the broadcast. If not applicable, on the other hand, we come to point X im Scheme of operation continued in Figure 5B.

As indicated in FIG. 5B, the next step is to go to, the first consonant in the word in register 12 search, starting from the left, if no consonant of column 12 of FIG. 4 is found in the word, then the code is FFFE, according to the entry in column 12 and the coding in column 11, in the case of a vowel ending with a code E, as shown in column 13. The resulting hexadecimal code is used at point Y of the flow chart in Figure 5C, discussed below.

If a consonant is found, the corresponding first hexadecimal code character (called "nibble" in the working scheme) is found in column 11, corresponding to the consonant found in column 12 of FIG. The next step, as indicated in the flow chart of Figure 5B, is to search for the second consonant. If a second consonant is not found, then the second and third hexadecimal characters (nibbles) are FF corresponding to the character mentioned in column 11 of FIG. H in the event that a consonant is not found in the list given in column 12.

1 3 0 0 1 B / 0 7 2 2

If a consonant is found in the list in column 12, the second hexadecimal code character is the corresponding character found in column 11. The next step is in looking for a third consonant. If there is no third consonant according to the one in column 12 is found, the third hexadecimal character is F, as can be seen in columns 11 and 12. Will on the other hand, if a third consonant is found according to the list in column 12, the corresponding hexadecimal is located Character in column 11 to be the third character of the address to build.

If the second and third consonants are not found on the list in column 12 or if the third consonant is not found, or if the third consonant is found, three of the hexadecimal characters are now one address consisting of four hexadecimal characters. To add the fourth hexadecimal character or "nibble" the last four letters of the word are compared with each of the four character endings in column 4. Will If a match is found, the fourth character is derived from the corresponding hexadecimal character from column 13 - If a match is not found, a comparison is made for the three-character endings and then a comparison for the two-character endings. The successive steps are not specifically shown in FIG. 5B for the sake of simplicity. Finally, a comparison is made for the one-character endings in column 14

130015/0722

or for a vowel ending or for an ending that cannot be found in column 14. In the case of a vowel ending, this is Coding for the fourth character a hexadecimal E, and if no match is found, for any 14, the hexadecimal coding is an F. Accordingly, it is now a four hexadecimal character Address formed so that one can now proceed to point Y of FIG. 5C in the working scheme.

Referring to FIG. 5C, the code comprising four hexadecimal characters is now used to search for a word in the PROM 20. The address in the PROM is accessed and the word is extracted character by character until an EOR is detected and the results are stored in comparator 22 of FIG. 1 for comparison with the word in register 12. If identity is established, the word is found, which is indicated by answering "yes" to the question "found?" in is displayed in the work scheme. The word then becomes temporary Storage for the program or for the program entered directly, as indicated in the operating diagram of FIG. 5C is, after which the input is made at point Z in this case. If the word doesn't indicate identity, the comparator is and a one is added to the hexadecimally encoded address. If the original Address is incremented so many times that the search limit set for the vocabulary in question is exceeded we move on to point W of the flow chart of FIG. 5A, and it becomes the following

130015/0722

Worfcis characters encoded as ASCII code, with EE8O or EE9O precedes or follows. If on the other hand the limit has not been exceeded, a loop occurs, as indicated in the work scheme, in which the word is checked again with the incremented address. This process continues until either Store the word identical to the word found in the register or the search limit is exceeded; thereafter the characters of the word are sent or transmitted in the ASCII code.

In all cases - be it that the tested word is actually- ' borrowed in the PROM or not - causes the mixing program a translation of the word into an address. If, according to the result of a check, the address is actually the contains the word in question, the address is sent or saved for broadcast. If the address does not contain the word to which the search extends, an incremented address can be checked for identity and if a if such is found, the incremented address is sent or saved for the broadcast. In each such case, a unique address will be sent to the recipient, who finds the word at the address. If the word is not at a start address or at an incremented address will be found within the limits of the incrementation the letters of the word are sent as ASCII characters.

1 3001 S / 0722

ORIGINAL INSPECTED

- / ft -

A problem associated with the system described here is that the encoded address often points to another Letters in the word indicate than its initial letter. For example, consider the two words "mares" and "marsupial" under the algorithm as specifically described using the tables of FIG we mean that "mares" encodes as address 7ABC, while "marsupial" encodes as 7ABF. When the two words are in a memory, as shown in Fig. 6, obviously different program loops have to be run through, before the word "marsupial" is found at 7AC2. In addition, some words obviously have to have a lot Loops are looped through, which takes a considerable amount of time, before the first letter of the desired word found and the identity is recognized. This fact requires a realistic program to run a high limit as the maximum number of loops for incrementation in the programming device.

Much time can be saved by using indirect addressing. The PROM 20 will work on a pair of Stores reduced or split into two stores, one of which is a 6 4K store i.e. 2EXP16, which is at each digit has a four-digit hexadecimal number for use representing an address. The 64K Spelcher (which can be an 8-bit by 8-bit field) stores addresses, which are always the starting addresses of a word of four hexadecimal digits are as illustrated in FIG.

130015/0722

-3T-

Each of the addresses in the note PROM points to the first letter of the corresponding word stored in the vocabulary PROM 20. In effect, the hint completes or auxiliary memory incrementing or advancing of addresses, if necessary, but not always, by 1. From this example it can be seen that an excessive number of repeated loops in programming is avoided or reduced by the aid of indirect addressing.

Coding examples which reduce the number of bits to be transmitted by using the method and of the device according to the invention are illustrated in Figs.

Fig. 9 contains in column I in plain text the message "more information dear Stanley, surprised? ". Column II shows the corresponding ASCII code in hexadecimal characters with a pair of hexadecimal characters (or a byte) for each alphabetic character is. In the next column III the ASCII code is expressed in binary digits or bits. In of the following column IV are the characters of the plain text according to the invention expressed in hexadecimal characters, and . in column V in bits. There are 80 hexadecimal characters for the coding of the plain text according to the ASCII code or 320 binary digits available. But there are only 28 hexadecimal characters

130015/0722

. 03 ·

or 112 binary digits required using the invention. It can be seen from this that a code compression of almost 3: 1 can be achieved by using a coding according to the invention.

Another example is explained in FIG. The plain text here is "the hashing algorithm serves to convert any word in the system's vocabulary into a four nibble (16 binary digit) code ". If this message is after the ASCII code encoded, it requires 119 bytes or 952 binary digits. By using the coding algorithm according to the invention, only 54 bytes, or 432 binary digits, are obtained for the same Save process. In this example there is a saving of about 2: 1 compared to the ASCII code has been achieved.

1300 15/0722

Claims (22)

Claims Method for the transmission of words, each of which is composed of characters which are encoded in a specific binary digit code _T characterized by: that when entering a word to be transmitted, the word input is coded character by character in the code mentioned will, that from this word code a second code, which serves as a memory address, is generated, that the address is entered into a first memory to the word stored at the address in the memory derive therefrom that the stored addressed word is compared with the entered word for identity will and that the address is sent if and only. if identity is established during comparison, 13001S / 0722 ζ- ■ '■■■ ■ "■ - ■■ that the sent address received, the received address in a second memory, the the same words are stored at the same addresses as the first memory contains the word stored at the address entered in the second memory is read from the second memory or is issued. 2. The method according to claim 1, characterized in that the codes are binary digit codes. 3. The method according to claim 1, characterized in that the second code is generated from the first code by a mixing algorithm. 4. The method according to claim 1, characterized in that then and only if no identity is established, a unique character that is a broadcast character for characters means, generated and then the word in the memory is sent character by character and thereby the Sending any word from the memory is prevented. 5. The method according to claim 4, characterized in that a second unique character is also generated, which means the end of the broadcast, one character at a time. 130015/0722 6 .. The method according to claim 1, characterized in that ^: that a unique character, which means broadcasts character by character, is generated if and only if identity is not determined in the comparison, that the waiting in the register is transmitted character by character until all characters in the register have been transmitted that a second unique character is generated and transmitted, which indicates the termination of the transmission Character for character means and that the characters transmitted in this way receive characters will. 7. Delivery system for messages consisting of words, each of which is composed of characters- which are coded in a specific code, characterized by the following features: a register for receiving the input of a word to be transmitted, this input character by character is coded in a specific code, code generating means for generating an address from the word entered in the register, a memory in which addresses as well as words to these Addresses are stored, which is switched so that it receives the generated addresses and that at each outputs stored words at an address, 130015/0722 ORIGINAL INSPECTED means for comparing the word output from the memory with that received in the register Word on identity of both, means for transmitting the memory address of the word if and only if the identity comparison means the identity affirmed, a device for receiving the sent address with a second register, a second memory that has the same addresses and the contains the same words stored at these addresses as the first memory and which is connected in such a way that it receives the transmitted address from the second register and the word stored at this address issues. 8. System according to claim 7, characterized in that the Code generator means a device for executing a mixing algorithm for the determination of addresses of words in the word code. 9. System according to claim 7, characterized by the following further features: a device for recognizing one or more characters of a word entered in the first register, a device for generating a unique character as a function of the recognition, and a device that responds to this unique character and is used to record the transmission or transmission 13001 B / 0722 direction to change the characters of the word entered in the first register character by character to submit. 10. System according to claim 9, characterized in that the generating means further includes means for generating a second unique character, which serves to indicate the completion of the transmission of characters. 11. Procedure for transmitting or sending words, each of which is composed of characters encoded in a first code, characterized by; that a word to be transmitted is entered and encoded character by character in the first code, that a second character, which is used as a memory address, is generated from the code entered in this way, that is to say the address is entered in a memory to derive the word stored at the address in the memory, and that this address is transmitted or sent. 12. The method according to claim 11, characterized in that the code is a binary code. 13. The method according to claim 12, characterized in that one or more characters of a word when entered character by character with the meaning of a transmission are recognized and dependent on this 13001 δ / 0722 Character recognition the entered word character by character is transmitted. 14. The method according to claim 13, characterized in that a second unique character is additionally generated, which corresponds to the termination of the transmission character by character. 15. The method according to claim 11, characterized in that in the first code, the word in the memory at the address for identity with the word to be transmitted before transmission the address is compared and the address is only transmitted if identity has been established will. 16. The method according to claim 15, characterized in that if no identity is established, the address is incremented, the incremented address to derive the word stored at the incremented address entered into the memory and the word removed from the memory again compared with the word to be transmitted will. 17. The method according to claim 16, characterized in that the incrementation is canceled and the comparison process repeats until identity is established or a maximum number of incrementing operations is performed has been. 13001 S / 0722 18. The method according to claim 17, characterized in that if no identity is established, the word in the first code is transmitted. 19. Message transmission system for the transmission of messages, which consist of words, each of which is composed of characters encoded in a code are indicated by a register for receiving the input of a word to be transmitted, which encodes character by character in the code is, a code generating device for generating an address from the word entered in the register, a memory which contains addresses and words stored at these addresses and is connected so that the The generated address is received and the corresponding word stored at the address is output,. means for comparing the identity of the word output from the memory and that in the register received word and means for communicating the memory address of the word if and only if the identity comparison means Indicates identity. 20, system according to claim 19, characterized in that the code generating device has a device for implementation a mixing algorithm for determining addresses from words in the word code. 13001S / 0722 21. System according to claim 19, characterized by the following additional features: a device for recognizing one or more characters of a word entered in the register, a device for generating a unique character, depending on the recognition and a device that serves, depending on the unique character, the transmission or transmission device to cause characters of the word entered in the register to be transferred character by character. 22. System according to claim 21, characterized in that the generating device additionally contains a device for generating a further unique character, which characterizes the termination of the transmission for indicating signs. 13001S / Q722